writing-skills

Writing agent skills with a test-driven development approach involves enforcing a RED-GREEN-REFACTOR workflow on documentation, ensuring tests precede guidance and rationalizations are captured for reliable deployment.

Test-driven development for skill documentation is a process that applies RED-GREEN-REFACTOR cycles to authoring, requiring YAML frontmatter, rationalization tracing, and progressive disclosure structure to validate AI agent skills.

Yes, creating agent skills requires YAML frontmatter with name and description fields. This structured metadata is necessary to validate the skill documentation before deployment across AI agents.

The best way to structure agent skills is using a SKILL.md template with standard sections, YAML frontmatter, and progressive disclosure structure, validated through pressure scenarios and rationalization tables.

Skill verification requires rationalization tracing to capture the reasoning behind documentation decisions, ensuring that the progressive disclosure structure and testing patterns remain reliable across AI agents during deployment.